Methylene Derivatives as Intermediates in Polar Reactions. VIII

Since chlorodifluoromethane reacts with methanolic sodium methoxide 150 times as fast as chlorofluoromethane does at. 35° despite the fact that -fluo...
0 downloads 0 Views 530KB Size
DIFLUOROMETHYLENE IN THE REACTION OF CHLORODIFLUOROMETHANE 5493

Oct. 20, 1957

temperature was kept under 35’ by external cooling. When the reaction became sluggish, more catalyst was added. Reduction was complete when a viscous yellow oil appeared and only a faint odor of nitrobenzene could be detected.” Water was added and the mixture placed in an icesalt-bath. The yellow oil soon crystallized. The crude azoxybenzene was removed by rapid filtration. Some of the impurities oiled out leaving behind a yellow solid. The oil was chilled and the process repeated for an additional yield of crude azobenzene. The azoxybenzene was recrystallized from alcohol, yield 5.5 g. (6570), m.p. 34-35:. On a second recrystallization the product melted a t 35-36 . 2,2’-Dimethy1azoxybenzene.-To 10 g. of o-nitrotoluene (0.073 mole) was added 9 ml. of lOOYo hydrazine hydrate (0.18 mole) and 20 ml. of 95% ethanol. Some Raney nickel catalyst was introduced. There was much effervescence and the mixture turned slightly green, then yellow. The temperature of the reaction was kept under 35”. When

the reaction became less vigorous, the solution was orange. More catalyst was added and the mixture allowed to stand a t 0” overnight. The resulting orange-yellow crystals of 2,2’-dimethylazoxybenzene were collected by filtration and recrystallized from a mixture of alcohol and water. A second crop was obtained by adding water to the mother liquor, from the original filtration and treating the resulting dark orange-red oil with dil. HC1 until it crystallized. The total yield was 3.2 g. (39%), m.p. 57.5-58.5”.18 4,4‘-Dimethy1azoxybenzene.-Ten grams of p-nitrotoluene (0.073 mole) was suspended in a solution of 9 ml. of 1 0 0 ~ ohydrazine hydrate (0.18 mole) and 30 mi. of 9570 ethanol. Some Raney nickel catalyst was introduced and the mixture stirred for several hours. The p-nitrotoluene gradually dissolved and yellow crystals of 4,4’-dimethylazoxybenzene were deposited. When effervescence had ceased, the product was filtered and recrystallized from a mixture of alcohol and water. This yield was 3.4 g. (427,), m.p. 7O-7l0.l8

(17) If the reaction was allowed to proceed too far, an orange oil or a low melting solid was obtained which was impossible to purify by recrystallization.

STASFORD,

and

(I8) L’

’‘

Ron’ A n n ” 4681

(1929)‘

CALIFORSIA SANFRANCISCO, CALIFORNIA

[CONTRIBUTION FROM T H E SCHOOL O F CHEMISTRY O F T H E

GEORGIA INSTITUTE

O F TECHNOLOGY]

Methylene Derivatives as Intermediates in Polar Reactions. VIII. Difluoromethylene in the Reaction of Chlorodifluoromethane with Sodium Methoxidel BY JACK HINEAND JOHNJ. PORTER RECEIVEDMAY9, 1957 Since chlorodifluoromethane reacts with methanolic sodium methoxide 150 times as fast as chlorofluoromethane does at 35” despite the fact that a-fluorine substituents are known to decrease S Nreactivity, ~ the reaction of chlorodifluoromethane is very probably not S N in~ character. It is instead believed to involve an initial or-dehydrohalogenation to give difluoromethylene, a reactive intermediate that yields a mixture of methyl difluoromethyl ether and trimethyl orthoformate. The intermediate, methoxyfluoromethylene, appears to be involved in the production of the orthoformate. The a-elimination mechanism is further supported by the observation that while chlorodifluoromethane reacts only very slowly with sodium thiophenoxide alone, in the presence of sodium methoxide the reaction (which yields phenyl difluoromethyl sulfide) is relatively rapid. The thiophenoxide ion is found to be more than one hundred times as effective as methanol a t combining .vvith difluoromethylene.

Introduction Chloroform and several other haloforms have been shown to undergo basic hydrolysis by the mechanism2, CHXs

+ OH-

+ H,O + x-

CXB-

cx3- +cxy

followed by rapid reactions of the intermediate dihalomethylene. This mechanism represents the principal reaction path only because such possible alternatives as the S K mechanism4 ~ are considerably slower. Since a-fluoroalkyl halides are considerably more reactive by the S Nmechanism ~ than are the analogous chloro, bromo and iodo derivat i v e ~ it , ~ was of interest t o learn whether the di(1) Ref. 7 is considered part VII. For part VI under a somewhat different series title see J. Hine, N. W. Burske, M. Hine and P. B. 79, 1406 (1957). Much of the content of Langford, THIS JOURNAL, parts VI11 and IX was presented in a talk a t the Sixth Biannual Conference on Reaction Mechanisms, Swarthmore, Pennsylvania, September 13, 1956. (2) J. Hine, ibid.. 72, 2438 (1950); J. Hine and A. M. Dowell. Jr., ibid., 76,2688 (1954). (3) J. Hine, A. M. Dowell, Jr., and J. E. Singley, Jr., ihid., 78,479 (1956). (4) For the meaning of the term “Sh-2” see J. Hine, “Physical Organic Chemistry,” McGraw-Hill Book Co., Inc., New York, N. Y., 1956, Chap. 5. (5) J. Hine, C. H. Thomas and S. J. Ehrenson, THIS JOURNAL, 71, 3886 (1955); J. Hine, S. J. Ehrenson and W. H. Brader, Jr., ibid., 78, 2282 (1956).

halomethylene (or a-elimination) mechanism is operative for such haloforms as chlorodifluoromethan e. Results and Discussion The Reaction of Chlorodifluoromethane with Sodium Methoxide.-The only organic products observed in the reaction of chlorodifluoromethane and sodium methoxide in methanol were trimethyl orthoformate and methyl difluoromethyl ether. Although the difluoromethyl ether, a new compound, was not analyzed because of its tendency to decompose on standing, the following evidence exists for its structure. I t s method of preparation is analogous to that used for ethyl difluoromethyl ether6 and isopropyl difluoromethyl ether.’ il survey in Lange’s “Handbook of Chemistry” of twenty-seven series of methyl, ethyl and isopropyl compounds showed t h a t the difference in boiling point between the methyl and ethyl compounds was 3.0 to 12.0’ (av. 8.7 + 2.6) greater than between the ethyl and isopropyl compounds. In view of this the boiling poipt of methyl difluoromethyl ether (-4’) is quite plausible considering the values reported for the ethyl (23.7’) and isopropyl (44.5’) analogs. Since an alkoxy substituent is known to activate a -CF,- group so t h a t its acid hydrolysis (6) A. L. H e m e and M. A. Smook, ibid , 72, 4378 (1950). (7) J. Hine and K. Tanabe, ibid., 1 9 , 2654 (1957).

5494

JACK

HINEAND

t o a carbonyl group occurs readily,6 we carried out the sulfuric acid-catalyzed hydrolysis of our material and identified methyl formate as a reaction product. Our structural assignment is also supported by the infrared spectrum of the compound and by stoichiometric considerations. The methyl orthoformate and methyl difluoroiiiethyl ether isolated accouiited for SOY0 of the haloform t h a t reacted (as measured by the amount of sodium chloride formed) in a synthetic experiment despite t h e fact t h a t some of the volatile ether was undoubtedly lost. Assuming these arc the only organic products, their relative yields may be calculated from the concentritions of metfioxide ion used up and of chloride ion formed. Thus so

=

A[[?rleO-]

- 2 [ c F l-

1 9

(1)

JOHN

J. PORTER

Vol. 79

activity,j chlorodifluoroinethane should react even more slowly if its mechanism were S\2. Since i t instead reacts more than 150 times as fast, the S,2 mechanism seems implausible. The reaction thus appears t o be initiated by an a-elimination with the formation of a dihalomethylene8 The unreactivity of alkyl halides in general and of fluoroform in particular37 shows t h a t dehydroiluorznatzon is probably only a minor side reaction. Hence the major reaction must be difluoromethylene formation. Some of the intermediate difluoroinethylene adds methanol to yield nletliyl difluorometbyl ether and some reacts to give methyl orthoformate. The fact t h a t fo, the fraction of otthofoLmate,doss not increase dcring a run showc-s that the difluoromethq.1 ether does not react fux theit o give orthoformate. The ortllocster is tlierefcrc, very probably formed through the intermediate, n~ethoxyfluoroniethylenc.~ Thus vie arrive 'it the reaction scheme

CHClI'L + CH30- ---f Cl'z -T- CHaOII + i ' l wherej, is tlie fraction of tlie reacted haloforiii t h a t gave methyl orthoformate. The analogous fracCFZ ---P CHjOCHF2 tion in the reaction of potassium isopropoxide with chlorodifluoromethane was found to increase somewhat with increasing alkoxide ion ~ o n c e n t r a t i o n . ~ This apparently is true in the present case since a CHaOCF --j. (CH3CI)aCH much higher fraction of methyl orthoformate was produced in t h e synthetic run, using concentrated Capture of Difluoromethylene by Thiophenoxide sodium methoxide, than in the kinetic runs. Never- Ions.-In order to bring a different sort of evidence theless, i t was found t h a t fo does not change greatly to bear upon the reaction mechanism, we attempted during a kinetic run a t the reactant concentrations to capture the intermediate difluoroniethylene by we used. Assuming fo constant, the following ki- use of the thiophenoxide ion, a reagent found previnetic equation may be derived. ously to be effective a t combining with dihalomethy l e n e ~ . ~We . ~ first studied the reaction of chlorodifuoromethane v i t h sodium thiophenoxide alone. where n = JCHClF2]o, b = [MeO-]o, x = A- This was found to be a very slow process a t 3.5' in [CHClE',]t, t = time and k is expressed in I. mole-I methanol, the second-order rate constant being 1. mole-' set.-'. This small reac(0haloform) set.-". Typical kinetic data calcu- about 5 X tivity tolvard such a powerful nucleophilic reagent lated from this equation are shown in Table I. The fall in rate constants with time probably is due to is itself added evidence that the 150-fold faster rc~ character. the loss of the volatile haloform (b.p. -40") from action with methoxide ion is not S S in ~ the thiophenoxide ion the reaction solution, so t h a t perhaps the value since in other S S reactions by ex- has been found t o be from twenty to ten thousand should be estimated as about 7.S X times as reactive as the rnethoxide ion.5 Conclutrapolation t o zero time. sive evidence against the Sh-2 mechanism, however, TABLE I was found in the observation that the reaction with THE REACTIOSO F CHLORODIFLUOROMETHANE WITH S O thiophenoxide to give phenyl difluoromethyl sulDIUM M E T H O X I D B I N M E T H S S O L AT 35" fide was powerfully catalyzed by sodium methoxide. The presence of sodium methoxide caused the thio[CHC1F2lo=0.03721; [SaOMeIo = 0.02918; fo =0.075 Time, 10'k, phenoxide to react as much as 60 times as fast as i t [ M e 0 -1 1. mole-' set.-' sec. would have in the absence of added base. Phenyl 7595 0.023232 7.58 difluoromethyl sulfide was isolated in more than 14478 .0!918 7.84 GOYo yield i n one run in the prescnce of added so29910 ,014m 7.08 dium met.lioxide. 32270 .01:iA3 7.06 JVe have made a quantitative estimate of the 6094.5 ,00924 6.30 tendency of the thiophenoxide ion to conibine with 75802 .(!0707 6 A9 difluoromethylene as follows. Difiuoroiiiethylme A\.. 7. 09 f 0.1 1 might be expected t o combine with any of the thrcc I n order to estimate how fast methoxide ion nucleophilic reagents. inethaiiol, rnethoxide ion or should react with chlorodifluoromethane by the thiophenoxide ion. SeglcctiIig t!i:it sniall though SN%mechanism, we determined the approximate significant part of the reactioii that Ijroceeds far (S) For evidence t h a t this a-elimination proceeds b y a concerted rate constant for the reaction of chlorofluoromethmechanism, not involving a trihalomethyl carbanion, see J. Hiiie ant1 ane with sodium niethoxide. This rate constant P. B. Langford, ibid., 7 9 , 5497 (196i). was found to be about 4.5 X 1. mole-' sec.-l (9) T h e argument t h a t this is so is analogous to that given brieAy and since a-fluorine is known to decrease Sx2 re- in ref. 7 and to be presented in more detail in a subsequent article.

DIFLUOROMETHYLENE I N T H E REACTION OF

Oct. 20, 1957

enough to liberate fluoride ions, we may write

+- CHaOH 2CHaOCHFz kB CF2 + CHIO- +CHaOCHFz CH30H CF2

and the rate of reaction of methoxide ion is expressed

By the use of both acidimetric and iodimetric titrations the changes in thiophenoxide and methoxide ion concentrations may be determined. We now simplify eq. 4 by neglecting the reaction of difluoromethylene with methoxide ion, relating later how this neglect is justified. With this simplification eq. 4 becomes

In Table I1 are shown data on a run, calculated according to eq. 6 which is based on the assumption TABLE I1 REACTIONO F CHLORODIFLUOROMETHANE WITH SODIUM T H I O P H E N O X I D E AND SODIUM METHOXIDE AT 35""

Q

[CaHsS-I

135 0.1509 ,1304 2985 ,1153 8235 .lo22 21883 ,0997 35790 70530 .0970 [CHClFzlo 0.1388.

[MeO-I

k,/k,

0.2043 4.55 .1722 4.55 .I452 .I233 4.81 4.13 .lo40 .lo16 _ _ _4.30 _ .4v, 4 47

5495

The numerous reports of difluoromethylene as an intermediate in gas phase reactions'O certainly strengthen the evidence t h a t it and other dihalomethylenes are intermediates in the hydrolysis of halof orms.

Experimental

From this reaction scheme (treating k M as a firstorder rate constant) and the assumption that difluoromethylene is formed only by the action of methoxide ion on the haloform, we may express the rate of reaction of thiophenoxide ion

Time, sec

CHLORODIFLUOROMETHANE

* 0 20

that thiophenoxide ion is competing only with the solvent methanol for the intermediate difluoromethylene. If the methoxide ion had been a major competitor, the calculated values of k T , / k M should increase as the reaction proceeds and the concentration of methoxide ion decreases. Since methanol is present a t a concentration of about 25 M , the values of kT ' k M found show t h a t the thiophenoxide ion is more than 100 times as reactive.

Reagents.-Chlorodifluoromethane (Matheson Chemical Co.) and chlorofluoromethane (du Pont) were used directly from their cylinders without further purification. Infrared measurements showed t h a t the chlorodifluoromethane contained less than 2% dichlorofluoromethane and less than O.5Y0 fluoroform. Methanol was purified by the method of Fieser, using magnesium." The thiophenol (Matheson) used was found by iodimetric titration t o be 99% pure. Turmeric yellow, rosolic acid and p-nitrophenol were all found to be satisfactory indicators for the titrations carried out in methanolic and aqueous methanolic solutions. Products of Reaction of Chlorodifluoromethane with Sodium Methoxide.-Seventy-two grams (3.1 moles) of sodium was dissolved in about 600 ml. of methanol in a oneliter three-neck flask equipped with a fritted-glass gas-inlet tube and a -80" reflux condenser. Chlorodifluoromethane was then admitted t o the flask rapidly causing an exothermic reaction. After 90 minutes the continual return of cold product from the reflux condenser had caused so much cooling t h a t the reaction was proceeding rather slowly. Therefore the admission of haloform was interrupted and after allowing a few minutes for the residual haloform to react and the Dry Ice in the condenser t o sublime the reaction flask was heated until methanol began to reflux, the more volatile products being collected in a trap a t -80". The admission of chlorodifluoromethane was resumed until the reaction appeared to be essentially complete and the distillation of volatile material repeated. During the reaction a large amount of salt had precipitated from the reaction mixture. This was collected on a filter, washed with methanol, dried and found to weigh 130 g. Titration for chloride (Fajans) and fluoridelZ and for base showed t h a t the material included 87.0 g. (1.49 moles) of sodium chloride and 42.4 g. (1.01 moles) of sodium fluoride and 0.01 mole of base. The methanolic filtrate was fractionally distilled through a 30-inch column packed with glass helices giving as the only products methanol containing some trimethyl orthoformate and pure trimethyl orthoformate (b.p. 98-99.5" at 740 mm.-infrared spectrum identical t o authentic material). Including t h a t material present in the methanol (analyzed by infrared measurements), 51 g. (0.48 mole) of trimethyl orthoformate was obtained. The volatile material collected at -80' was distilled through a 2 ft. column packed with glass helices and equipped with condensers through which refrigerated glycolwater was circulated. The fraction boiling a t -4' at 740 mm. occupied 54 ml. at -80', b u t i t was clear t h a t not all of the material wm being condensed by the coolant which was at -8". The material had infrared absorption maxima (in the order of decreasing intensity) at 9.53, 9.44, 8.97, 9.21, 8.18, 8.12, 8.25, 8.43, 8.84, 12.36, 12.21, 12.49, 7.29,7.23, 7.58, 6.84, 7.64, 10.77, (gas phase). This di6.89, 7.44, 3.43, 13.46 and 3.56 fluoromethyl ether, like the other two reported, is rather unstable, decomposing on standing to liberate hydrogen fluoride. When about 0.5 ml. of sulfuric acidowas added t o about 10 ml. of the material refluxing at -4 , there was a slow evolution of gas t h a t was not condensed in the -80" reflux condenser. Infrared measurements showed t h a t this gas contained carbon monoxide. The reflux condenser was then removed, the difluoromethyl ether allowed to evaporate and the remaining solution heated t o 50" with the evolution of gas. This gas, collected after passage through a short tube filled with sodium hydroxide pellets, had a n in(10) J. Dacey and J. Hodgins, Can. J. Research, 28B, 90, 173 (19.50); P. Venkateswarlu, Phys. Rev., 17, 676 (1950); R. K. Laird, E. B. Andrews and R. F. Barrow, Trans. Faraday SOL, 46, 803 (1950); J. L. Margrave and K. Wieland, J. Chem. Phys., 21, 1552 (1953); and references cited therein. (11) L. F. Fieser. "Experiments in Organic Chemistry," 3rd Ed., D. C. Heath and Co., Boston, 1955, p. 289. (12) I. M. Kolthoff and E. B. Sandell, "Textbook of Quantitative Inorganic Analysis," 3rd Ed., The Macmillan Co., New York, h'. Y., 1052, p. 721.

5496

JACK

HINEAND JOHNJ. PORTER

frared spectrum identical t o t h a t of methyl formate except for a band at 9.75 p , probably due t o silicon tetrafluoride. From data on related compounds the density of methyl difluoromethyl ether a t -80" is estimated as 1.10. Thus the yield of this product is 59 g. (0.72 mole) or 48'37' based on the haloform reacted (sodium chloride formed). Since a 325; yield of trimethyl orthoformate was obtained, go$;, of t h e haloform t h a t reacted was accounted for. From the relative yields of chloride and fluoride a 3470 yield of the orthoester and 6670 of the clifluoromethyl ether would be expected. Kinetics of the Reactions with Sodium Methoxide Alone. -Standard solutions of the organic halides were prepared by bubbling into methanol and weighing. The concentrations of several of the haloform solutions thus prepared mere checked by titration for chloride ion in reactions t h a t had proceeded well past 99yo completion. For both compounds kinetic runs were carried out by adding a known volume of standard sodium methoxide to a large volumetric flask containing the halide solution and then withdrawing samples for analysis at known times. From earlier experience on the reaction of bromofluoromethane with sodium methoxide5 i t was expected t h a t the methyl fluoromethyl ether formed initially in the reaction of CHzCIF might not react completely in the methanolic reaction solution b u t might hydrolyze if Tvater were added during the analysis of samples. T o minimize such hydrolysis, 10-ml. reaction samples were pipetted into 12 ml. of cold methanol and titrated quickly mith less than 12 nil. of standard aqueous acid. When more water was added to a titrated sample, acid was liberated revealing the presence of the fluoromethyl ether. R a t e constants were therefore calculated on the assumption t h a t only one rnethoxide ion is used per halide molecule. The value (4.5 i 0.6 X 1. mole-' sec.-lj was obtained a t 35'. Since chlorodifluoromethane was considerably more rcactive, its reaction samples Tvere addcd to a known volurnc of standard acid to stop the reaction. The excess acid was then back-titrated. Alppropriate experiments showed t h a t the difluoromethyl ether does not solvolyze appreciably in largely aqueous solutions during the course of an hour, apparently because of a decrease in Szl reactivity brought about by the second fluorine atom.I3 The titrations could therefore be carried out with aqueous solutions. For eight points, when the acidimetric titration was complete, the concentration of chloride was determined by the Fajans niethod. The value of f o calculated from eq. 1 ranged from 0.0922 t o 0.0672, decreasing slightly as the reaction proceeded. A n intermediate value of 0.075 was used in calculating the rate constants. Although fo probably varies somewhat Ivith the methoxide concentration, a n initial concentration of about 0.03 jl1 was used in all of our runs. The values of k obtained are not very sensitive t o the exact value of f o used, particularly when the haloform i.; in excess as in our reactions. Products of the Reaction of Chlorodifluoromethane with Sodium Thiophenoxide.-One hundred twenty-fire grams (5.4 molesj of sodium was dissolved in about 800 mi. of methanol in a one-liter three-neck flask equipped with a stirrer, Dry Ice cooled reflux condenser and a gas dispersion tube. T o the flask was added 112 g. (1.02 moles) of thio(13) Cf.J. Hine and D. E. Lee, THISJOURNAL. 74, 3182 (1952).

VOl. 79

phenol with stirring and then chlorodifluoromethane was bubbled into the reaction mixture. A t three times during the 7 hr. of reaction, the gas addition was discontinued, a sample from the reaction flask titrated acidimetrically and iodimetrically to follow the progress of the reaction, and the flask m.s ivarmed t o remove the difluoromethyl ether t h a t had formed. TT'hen all of the sodium niethoxide had been destroyed, 0.10 mole of thiophenol remained unreacted. The completed reaction mixture was concentrated to about 300 ml. of a salt-liquid slurry by distillation and poured into about five liters of water. The heavy oily layer t h a t separated 'ivas taken up in methylene chloride and washed with alkali. It was then combined with a second methylene chloride extract of the tu-o aqueous layers and distilled until the methylene chloride xvas removed. Distillation of the residue yielded five fractions totaling 93.4 g. at pressures of 2 t o 7 mm. and temperatures of 29 to 50" and 25.1 g. of a sixth fraction boiling largely at 165-185° at 1-2 mm. Fractionation of about 30 g. from samples four and five gave essentially entirely material, b.p. 62-62" at 7 m m . , EZ5D 1.5084, d2541.2218, molecular refraction calcd. for CoH6SCHF: 38.69, found 39.11. Comparison of infrared spcctra and refractive indices showed t h a t practically all of the 93.4 g. was phenyl difluoromethyl sulfide (637' yield based on thiophenol that was used up). The phenyl difluoromethyl sulfide structure is further supported by the infrared spectrum. Absorption maxima (in the order of decreasing intensity) were found a t 9.32, 9.60, 13.55, 9.75, 14.60, 7.74, 9.14, 7.60, 13.06, 14.27, (3.99, 3.35 and 3.46 p. Distillation of the material at temperatures above about 60" gave a product t h a t became cloudy and red upon standing. This anoeared t o be due t o the txesence of acid and could be &;exited by storage over sodium carbonate or even by an alkaline wash. A n d . Calcd. for C7HsSFz: C, 52.48; H, 3.78; S, 20.02; F, 23.7%. Found: C, 52.20, 52.49; H,3.83, 3.83; S, 20.22, 20.12; F, 24.59, 24.78. Kinetics of the Reaction with Sodium Thiophenoxide.Standard sodium thiophenoxide solutions were prepared by the addition of a weighed amount of thiophenol to a known volume of standard sodium methoxide, thiophenol being used in slight excess to produce the "sodium methoxidefree" sodiium thiophenoxide. Runs were carried out like those with sodium methoxide alone except for the analvsis of the reaction samples. This was done by titration to the p-nitrophenol end-point (both the sodium methoxide and thiophenoxide are neutralized in this titration) and then determining the thiophenol by titration with methanolie iodine using its o v n yellow color as the indicator. The rate constant for the reaction of chlorodifluoromethane in the absence of sodium methoxide mas calculated from the second-order rate equation assuming t h a t one mole of thiophenoxide reacted per mole of haloform.

Acknowledgments.--The authors wish to express their gratitude to the National Science Foundation for a grant in support of this work and to E. I. du Pont de Neinours and Co. for a giit of chlorofluoromethane. ATLANTA,GEORGIA